Probabilistic inference of lossy links using end-to-end data in sensor networks

  • Authors:
  • Wei Zeng;Bing Wang;Krishna R. Pattipati

  • Affiliations:
  • University of Connecticut;University of Connecticut;University of Connecticut

  • Venue:
  • CoNEXT '07 Proceedings of the 2007 ACM CoNEXT conference
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Lossy links used in a sensor network affect network performance, and hence need to be detected and repaired [1, 2]. One approach to detect lossy links is that each node monitors the loss rates on its neighboring links and reports them to the sink. This approach, although straightforward, causes large amount of traffic. Another approach to detect lossy links is through end-to-end data that are transmitted periodically from sources to the sink(s) [3, 1, 2]. This end-to-end approach has the advantage of not generating any additional monitoring traffic. The challenge is, however, to develop accurate inference algorithms for lossy link detection based on end-to-end measurements.